News

27/05/2026
New paper accepted at PPSN 2026

Our paper "Inferring Structural Causal Models from Data with Grammatical Evolution", by Romina Doz, Francesca Randone, Luca Bortolussi and Eric Medvet was accepted at PPSN 2026! 

Understanding cause-and-effect relationships in complex systems is a fundamental challenge in science. While traditional statistics can reveal correlations, causal models go a step further by explicitly mapping how one variable causes another, and not just correlates to it.

In this paper, we tackle the following problem: can we automatically learn causal structures directly from data, without prior knowledge of the underlying system? Our approach couples evolutionary computation in the form of Grammatical Evolution (an algorithm able to generate programs mimicking natural selection) and probabilistic programming to represent causal models in a suitable form. 

We show that the approach is able to recover simple causal structures that then combine to form complex models. 

📖 Pre-print soon available
👾 Implementation available at: https://github.com/r-doz/Causal_Models_Synthesis

23/04/2026
Invited Seminar at TU Wien

As part of the joint FORSYTE/ISTA seminar series I held a seminar titled "Probabilistic Programming: Semantics and Statistical Perspectives".

22/12/2025
New paper accepted at TACAS 2026

Our paper "DeGAS: Gradient-Based Optimization of Probabilistic Programs without Sampling", by Francesca Randone, Romina Doz, Mirco Tribastone and Luca Bortolussi was accepted at TACAS 2026! 

Remember SOGA, our PPL to enable inference of probabilistic programs with continuous and discrete variables?

DeGAS is its evolution, aimed at solving complex optimization problems.

DeGAS augments SOGA syntax with parameters, and applies a semantic smoothing to ensure that the posterior is always differentiable, even in presence of discrete random variables. By doing that, it allows you to optimize the parameters using torch gradient-based optimization. We have used DeGAS to optimize parameters in CPS models observing very nice results. In particular, we were able to optimize complex loss functions, not just standard likelihoods. 

📖 Pre-print available at: https://arxiv.org/pdf/2601.15167
📦 Replication package available at: https://zenodo.org/records/18197807
👾 DeGAS available at: https://github.com/frarandone/DeGAS (repo under active development)